【文章內(nèi)容簡介】 recasting ? 假定系統(tǒng)中存在因果關(guān)系 過去 == 未來 ? 由于不確定性因素的存在使得預(yù)測很少準(zhǔn)確無誤 ? 對一組事物的預(yù)測比 對單個事物預(yù)測準(zhǔn)確 ? 當(dāng)預(yù)測時間跨度增加時， 預(yù)測精度將下降 我看這學(xué)期你 能得優(yōu)秀成績。 341 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 好的預(yù)測方法的基本要素 適時 精度 可靠 書面 342 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 預(yù)測進(jìn)行的步驟 步驟 1 確定預(yù)測目的 步驟 2 確定預(yù)測時間跨度 步驟 3 選擇預(yù)測技術(shù) 步驟 4 收集和分析數(shù)據(jù) 步驟 5 準(zhǔn)備預(yù)測 步驟 6 預(yù)測監(jiān)控 “預(yù)測 ” 343 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 預(yù)測的類型 ? 主觀判斷 – 主觀意見 ? 時間序列 – 使用歷史數(shù)據(jù)，認(rèn)為將來和過去相似 ? 聯(lián)合模型 – 基于自變量預(yù)測未來 344 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 主觀預(yù)測 ? 經(jīng)理人員的意見 ? 與顧客直接接觸人員的意見 ? 消費(fèi)者調(diào)查 ? 其它預(yù)測方法 –德爾非法 Delphi method 345 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 時間序列預(yù)測 ? 長期趨勢 –數(shù)據(jù)的長期移動 ? 季節(jié)性變動 – 數(shù)據(jù)短期規(guī)則變化 ? 不規(guī)則變動 – 異常情況引起的變動 ? 隨機(jī)變動 – 各種可能性引起的變動 346 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting Ch 10 7 169。 2022 by PrenticeHall Inc Russell/Taylor Oper Mgt 3/e 需求變動的形式 需求 時間 趨勢變動 隨機(jī)變動 需求 時間 季節(jié)變動 需求 時間 需求 時間 周期變動 帶季節(jié)性的趨勢變動 347 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting ? 簡單易用 ? 基本上不需要成本 ? 不需要數(shù)據(jù)分析 ? 容易理解 ? 不能提供較高的精度 ? 能夠作為衡量精度的標(biāo)準(zhǔn) 簡單預(yù)測法 348 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting ? 平穩(wěn)時間序列數(shù)據(jù) –F(t) = A(t1) ? 季節(jié)變動 –F(t) = A(tn) ? 長期趨勢數(shù)據(jù) –F(t) = A(t1) + (A(t1) – A(t2)) 簡單預(yù)測法應(yīng)用 349 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 簡單預(yù)測 唉 , 給我點(diǎn)時間 .... 上周我們賣了 250輪胎.... 那么 , 下周我們應(yīng)該賣 .... 350 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 平均方法 ? 移動平均 ? 加權(quán)移動平均 ? 指數(shù)平滑法 351 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 簡單移動平均 圖 34 MAn = n Ai i = 1 ? n 353739414345471 2 3 4 5 6 7 8 9 10 11 12實(shí)際數(shù)據(jù) MA3 MA5 352 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting Ch 10 14 169。 2022 by PrenticeHall Inc Russell/Taylor Oper Mgt 3/e 加權(quán)移動平均 ? 可調(diào)整和反映簡單移動平均法中不同時期數(shù)據(jù)的影響 WMAn = i = 1 ? Wi Di 此處 , Wi = 第 i 期的權(quán)重，其百分值在 0 ~ 100 之間 ? Wi = 353 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting Ch 10 15 169。 2022 by PrenticeHall Inc Russell/Taylor Oper Mgt 3/e 加權(quán)移動平均之例 月份 權(quán)重 數(shù)據(jù) 八月 17% 130 九月 33% 110 十月 50% 90 十一月的預(yù)測 3 130 50 90 0 33 110 0 17 130103 4WMA W Di ii??? ? ???( . )( ) ( . )( ) ( . )( ). 354 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 指數(shù)平滑法 ?假設(shè) —最近的觀測具有最高的預(yù)測價值 . – 因此在預(yù)測時，應(yīng)該給更近的數(shù)據(jù)賦予更高的權(quán)重 . Ft = Ft1 + ?(At1 Ft1) 355 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 時期 實(shí)際數(shù)據(jù) A l p h a = 0 . 1 誤差 A l p h a = 0 . 4 誤差1 422 40 42 2 . 0 0 42 23 43 4 1 . 8 1 . 2 0 4 1 . 2 1 . 84 40 4 1 . 9 2 1 . 9 2 4 1 . 9 2 1 . 9 25 41 4 1 . 7 3 0 . 7 3 4 1 . 1 5 0 . 1 56 39 4 1 . 6 6 2 . 6 6 4 1 . 0 9 2 . 0 97 46 4 1 . 3 9 4 . 6 1 4 0 . 2 5 5 . 7 58 44 4 1 . 8 5 2 . 1 5 4 2 . 5 5 1 . 4 59 45 4 2 . 0 7 2 . 9 3 4 3 . 1 3 1 . 8 710 38 4 2 . 3 6 4 . 3 6 4 3 . 8 8 5 . 8 811 40 4 1 . 9 2 1 . 9 2 4 1 . 5 3 1 . 5 312 4 1 . 7 3 4 0 . 9 2指數(shù)平滑法舉例 356 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 選擇平滑常數(shù) 353739414345471 2 3 4 5 6 7 8 9 10 11 12Pe r i o dDemand? ? .1 ? ? .4 實(shí)際數(shù)據(jù) 357 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 線性趨勢方程 ? b ＝ 直線斜率 Yt = a + bt 0 1 2 3 4 5 t Y b = n (ty) t y n t 2 ( t) 2 a = y b t n ? ? ? ? ? ? ? 358 McGrawHill/Irwin Operations Management, Seventh Edition, by William J. Stevenson Copyright 169。 2022 by The McGrawHill Companies, Inc. All rights reserved. Forecasting 線性趨勢方程舉例 t y 周 t2 銷售 ty 1 1 150 150 2 4 157 314